University of Groningen
Marine benthic metabarcoding Klunder, Lise Margriet
DOI:
10.33612/diss.135301602
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Klunder, L. M. (2020). Marine benthic metabarcoding: Anthropogenic effects on benthic diversity from shore to deep sea; assessed by metabarcoding and traditional taxonomy. University of Groningen.
https://doi.org/10.33612/diss.135301602
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Download date: 24-06-2021
CHAPTER 1
General introduction
8 Chapter 1
The marine ecosystem is one of Earth’s most valuable natural resources. It provides us with food, medicines and valuable minerals, together termed: ‘ecosystem services.’ Marine ecosystems, next to the rest of the global environment, have been extensively altered by humans (Chapin et al., 2000). Habitat loss, overexploitation, invasive species, pollution and climate change are regarded as the main threats (Gray, 1997; Bijma et al., 2013). Even though human impacts are highest in the coastal zones, also the deeper seas are experiencing an intensification of anthropogenic influences (Merrie and Olsson, 2014). Without major changes in policy and our own behaviour, pressure on the marine environment will further increase with consequences for biodiversity.
Biodiversity, or biological diversity, has been defined in various ways (Gaston et al., 1996) and the shortest definition would be ‘the variety of organisms at all levels’ (Wilson, 1992). In general, a higher biodiversity is linked to a higher level of ecosystem functions (Törnroos et al., 2015).
But how can we measure biodiversity and how can biodiversity be linked to the state of an ecosystem? At this moment there is order-of-magnitude uncertainty in the estimation of species existing on earth overall, this uncertainty is even higher for marine (invertebrate) species (May, 1988; Cardoso et al., 2011; Costello, 2015) suggesting current methods are inefficient. This thesis explores a novel approach of next generation monitoring (molecular techniques) to measure marine benthic biodiversity and will compare this novel approach to the traditionally broadly applied morphological taxonomy approach for measuring benthic biodiversity.
The benthic ecosystem
The benthic ecosystem spreads from the intertidal beaches and estuaries towards the cold and dark deep seas and is mostly inhabited by invertebrates, a group of species often understudied (Cardoso et al., 2011). The benthic ecosystem in the intertidal zone and in coastal areas is generally characterized by high fauna abundancies and biomass whereas species diversity is often low. The number of species and biodiversity was estimated to peak around 500m depth (Costello and Chaudhary, 2017) with a decline again towards the deeper parts of the ocean. High concentrations of benthos in the deep sea can however be found around so called ‘hotspots’;
locations where chemicals seep from the ocean floor and chemotrophy gets a central role in primary production, for example in the surroundings of hydrothermal vents or cold seeps.
The importance of benthic communities has been widely recognized (Snelgrove et al., 1997;
Worm et al., 2006) because of their role in the break down and recycling of organic matter, bioturbation, and the food web (Levin et al., 2001; Thrush et al., 2006), not only in coastal systems but also in deeper areas and the open ocean. Moreover, benthic communities often hold key positions in the marine food web and changes in the environment are often rapidly reflected in the benthic community (Schmid-Araya & Schmid, 2001). Species composition can
9 General introduction
alter, or some species can completely disappear, under conditions of heavy stress (Balsamo et al., 2012). To be able to detect changes within the environment, measuring the benthic diversity at a temporal and spatial scale in a consistent and reliable way is required.
Taxonomy
The field of taxonomy, the scientific classification of living organisms, stands at the base for understanding biodiversity (McNeely, 2002; Thomson et al., 2018). The first attempt to classify organisms in groups was by the greek philosopher Aristotle in his book “The history of animals” written in 350BC. Although the system was not very methodological, it persisted for more than 2000 years until 1735. In this year, Linnaeus published his book “Systema Naturae”, the system of nature. Linnaeus developed the hierarchical system from kingdom to species and combined this with a binomial nomenclature. The framework for classifying living organisms described in Lenneaus is still, although greatly modified, essential for the system used currently. Within the current taxonomic systems, organisms are grouped in a taxonomic ranking system, consisting of eight taxa (Figure 1.1). The ability of naming species and clustering into groups is a first essential step in describing and comparing biodiversity.
Figure 1.1 | Present taxonomic ranking system including the major ranks: domain, kingdom, phylum, class, order, family, genus and species. The common cockle (Cerastoderma edule) is used as example.
Research slowly changed from exploration in the past centuries towards research more driven by empirical questions. Hence, scientific output changed from the discovery and description of species to estimating abundancies and biodiversity as a measure for ecosystem functioning.
To study benthic fauna, researchers have always relied on the morphological approach, which enables one to collect, identify and count species within a habitat.
1
10 Chapter 1
Morphological approach
Since Linnaeus, taxonomy has been defined based on morphological systematics. This system uses phenotypic characteristics to identify species. It assumes that individuals which are morphologically different represent different taxa and moreover, that closely related species differ less than distantly related species. Numerous dichotomous keys and pictorial guides have been developed to aid marine taxonomist to the identification of specimens up to the desired taxonomic level. By now, the world register of marine species (WoRMS, 2019) includes over 300 thousand marine animal species from oceans all over the world.
The collection of biodiversity data using the morphological approach has been generally accepted for marine benthic research (Figure 2). Within this method, the infauna and sedentary epifauna are collected by a sampling device of choice; often a grab or (box)corer (Flannery & Przeslawki, 2015).
Subsequently, the sample needs to be sorted; and sorted specimens will be identified to taxonomic levels. Specimens for each species can be counted and/or biomass can be calculated from ash- free dry mass (AFDM) measurements (Beukema & Cadee 1997; Compton et al., 2013) or by converting the blotted wet weight (Rutgers van der Loeff & Lavaleye, 1986; Daan & Mulder, 2005).
Figure 1.2 | Graphical abstract of the morphological approach including: collection of a marine sediment sample, species sorting and identification.
The morphological approach has been extensively used for decades and was the most prominent source in marine benthic biodiversity estimates. Limitations and biases of this method are well known and have been gradually accepted. Moreover, many limitations are introduced as an artificial boundary by researchers themselves. A prominent example is the division of species in size classes, i.e., fractionation of the sample. Because sorting a complete box core (up to 50 liters of sediment) and identifying specimens of all size classes is a time-consuming and nearly impossible, this process is preceded by a sieving or elutriation step. By sieving, the sample is divided into size fractions, the so called microfauna, meiofauna and macrofauna. Size classes are chosen arbitrary and differ depending on the research area and the researcher. Smaller size fractions are often discarded even though these fractions consist of species typically defined as good indicators for environmental health (Balsamo et al., 2012; Fraschetti et al., 2016). Meiofauna
11 General introduction
species, a size class often omitted, are the most abundant size class in the benthic environment and play an important role in the food web (Zeppilli et al., 2016).
Also, identifying the species in the bigger size fraction, the macrofauna, can introduce biases. Next to the well-known observer-bias, species can also be miss-identified due to miss-interpretation of cryptic species or the life-stage. Morphological taxonomy has become heavily opinionated as taxonomists differ in their assumptions in defining species (Hillis et al., 1974). The morphological approach is a time-consuming and costly process as it requires taxonomic expertise, which is scarce, particularly for invertebrates (Bucklin et al., 2010; Cardoso et al., 2011; Cowart et al., 2015). Taxonomy becomes more and more a rare skill and the fading of knowledge has been a widely addressed problem for decades (Wilson, 1985). Hence, biodiversity estimates are often impaired as only the ‘easily’ determined fraction, both for size and identification purposes, from the benthic biodiversity is considered.
Next generation biomonitoring
Morphological monitoring techniques focused primarily on measuring a defined part of biodiversity such as macrofauna species richness (Compton et al., 2003); marine benthic larvae (Gollner et al., 2016); or indicator species (Vandewalle et al., 2010). Measurements were often limited to such a defined assessment due to limitation in available techniques (Yoccoz, 2012).
In the past decade, a new set of tools has been introduced under the collective name: ‘next- generation biomonitoring (NGB) (Bohan et al., 2017; Makiola et al., 2020) including: remote sensing, citizen science, deep learning and molecular tools (Jackson et al., 2016). NGB allows for assessing (marine benthic) diversity in a more holistic way from genes to entire ecosystems. One of the molecular tools, DNA metabarcoding, will be explored within this thesis as a potential new standardized methodology for multiple taxonomic identifications (Taberlet et al., 2012). Although other NGB tools are not covered within this thesis, especially the combination of deep learning, or machine learning, in combination with DNA metabarcoding is highly potential for scaling-up both spatial and temporal resolution within biomonitoring programs (Cordier et al., 2019).
From DNA barcoding to the metabarcoding approach
In the past decades, there has been a growing understanding of genetics. The combination of the conventional DNA-sequencing (Sanger et al., 1977) and the concept of DNA barcodes (Hebert et al., 2003) made it possible to distinguish species based on their molecular code. These DNA barcodes are short variable gene regions containing valuable taxonomic information, flanked by more conserved regions which act as primer binding sites. The use of DNA removes the observer bias and opinion-based discussion about species identification. The DNA barcodes introduced by Hebert et al. (2003) focused at that time on a 658-bp region in the mitochondrial cytochrome c oxidase I (COI) gene of animals. Many more marker sites have been studied since then, and the
1
12 Chapter 1
18S rRNA gene is, next to the COI gene, most often used for marine benthic metazoan species (Deagle et al., 2014) (see Box 1).
Metabarcoding combines two technologies: high throughput DNA sequencing using NGS platforms and identification of species based on standardized DNA barcodes. Since the development of conventional DNA-sequencing by Sanger et al. (1977) DNA sequencing technologies have been impressively improved into the next-generation sequencing (NGS) techniques being introduced in the past decade. Next generation sequencing platforms uses massive parallel sequencing and can produce millions of reads for the barcode of choice within a run. This massive parallel sequencing allows us to identify many specimens at the same time from either a bulk or an environmental sample by comparing the obtained sequences with sequences of this barcode in publicly available reference libraries. Metabarcoding is by now a commonly used term for the study of the complete genetic material obtained from environmental samples (Taberlet et al., 2012). The metabarcoding workflow follows the same steps as general DNA barcoding: DNA extraction, PCR amplification, sequencing and taxonomic identification (Figure 3). However, where only a single species is targeted in DNA barcoding, metabarcoding targets the whole community. Therefore, metabarcoding allows us to assess biodiversity in a new, i.e.
consistent and replicable way across different ecosystems (Baird & Hajibabaei, 2012). Started in the microbiology, the different techniques are more and more used for higher taxa and other fields of biology, inter alia marine benthic biodiversity (Chariton et al., 2010 & 2015; Fonseca et al, 2010; Guardiola et al., 2015 & 2016; Lanzen et al., 2016; Sinniger et al., 2016).
Figure 1.3 | Graphical abstract of the metabarcoding approach including: collection, DNA extraction, PCR amplification, sequencing and taxonomic identification.
13 General introduction
Applications of metabarcoding
As with other novel methodological approaches, the introduction of a new method into the scientific world is faced by challenges. Although Linnaeus introduced his taxonomic system almost 300 years ago, changes to this system are still being suggested and implemented currently (Ruggiero et al., 2015; WoRMS, 2019). These challenges should be overcome by a mixture of meticulous testing as well as learning by trial-and-error. Although the term ‘metabarcoding’
was first introduced by Taberlet et al. (2012), a courageous study describing metabarcoding methods for marine ecological assessment was already published two years earlier by Chariton et al. (2010). Chariton et al. (2010) investigated if and to which extent metabarcoding sequence information could be used for an ecological assessment towards human impacts in a marine estuary. And even though this study used a very rough, unpolished version of the metabarcoding approach, the high potential of the metabarcoding approach was demonstrated. In the next couple of years, more of these ‘early’ adopters were able to demonstrate the potential of these methods for eukaryotic fauna in all kinds of ecosystems (i.a., Fonseca et al., 2010; Yu et al., 2012;
Thomsen et al., 2012).
However, at the same speed as these studies were published, critical reviews were published alongside in which limitations and challenges of these approaches were discussed widely (i.a., Coissac et al., 2012; Collins & Cruickshank, 2013; Cristescu, 2014; Pompanon & Samadi, 2015).
These critical reviews were essential as researchers were forced to think critically, spend time on much needed validations and become more creative in applying the metabarcoding approach.
The main challenges highlighted are the availability of reference sequences and the quantitative potential of the method. Gaps in the reference databases may lead to misidentifications or limited taxonomic resolution. Whereas the ability to acquire quantitative data can greatly enhance the power of ecological inferences and predictions (Gray, 2000; Yin and He, 2014), little validation studies have been published. Most validation studies focused either on the workflow for metabarcoding method solely (i.a., Brannock & Halanych 2015; Elbrech & Leese, 2015) or comparing metabarcoding with the traditional morphological approach in qualitative terms (i.a., Cowart et al., 2015; Lejzerowicz et al., 2015; Cahill et al., 2018). Numerous (sub)-methods have been developed by now and each ecosystem requires a specific experimental setup, therefore, it remains elusive which workflow is best suited for studying ‘the’ marine benthic ecosystems and to which extent biases and limitations of the workflow influence the outcomes.
1
14 Chapter 1
Box A: Primer test for marine benthic specimens
Identification of species through DNA methods is based on genetic variation of specific barcode regions. The sequences of these barcode regions ideally differ between species, to enable identification up to species level. As numerous copies of this specific barcode are needed for sequencing, a polymerase chain reaction (PCR) to amplify the specific region is key in the metabarcoding process. During this reaction, complementary oligo-nucleotides bind to the DNA and due to the copy activity of the polymerase enzyme, a new complementary string is build. To be able to sequence a variable barcoding region, it needs to be flanked by conserved regions - the so-called ‘marker’ sites - for an optimal binding of the primer oligo- nucleotides (Riaz et al., 2011). Next to being discriminant and having highly conserved regions, the ideal barcoding gene is a widely and intensively used gene, as the genetic region needs to be present in reference databases with sequences covering many taxa. In metabarcoding for animals, the mitochondrial cytochrome c oxidase subunit I (COI) and the nuclear ribosomal small subunit RNA gene (SSU rRNA), also called the 18S gene, are frequently used. Both marker sites have their advantages and disadvantages. The COI gene is generally highly discriminant, enabling OTU identification up to species level in many groups of organisms. The 18S gene is less variable and specimens can often only be identified up to family level. So, when high taxonomic recovery is needed, the 18S gene might me less suitable than the COI gene. On the other hand, the 18S gene contains more conserved marker sites enabling a better primer binding. Hence, the 18S can be applied in a more universal matter (Deagle et al., 2014). Six published primer pairs were tested in vitro for their ability to amplify the DNA barcodes from 35 different Wadden-Sea benthic species (Table 1). The 18S oligo-nucleotides designed by Hadziavdic et al., (2014) and Sinniger et al. (2016), performed best in terms of the ability to amplify all marine benthic species included in the test.
Table 1.1 | Oligo-nucleotide combinations tested for COI and 18S gene. For each combination, the score percentage out of 35 benthic species is shown.
Reference Gene Length Forward Reverse Score
Folmer et al., 1994 COI 658bp LCO1490 HCO2198 83 %
Lobo et al., 2013 COI 658bp LoboF1 LoboR1 50 %
Leray et al., 2013 COI 313bp mlCOIintF jgHCO2198 68 %
Hadziavdic et al., 2014 18S – V4/5 630bp F-566 R-1200 100 % Stoeck et al., 2010 18S – V4 375bp TAReuk454FWD1 TAReuk454Rev3 83 % Sinniger et al., 2016 18S – V1/2 450bp SSU_F04 SSU_R22mod 100 %
15 General introduction
Thesis outline
The aim of this thesis was to explore the application of molecular metabarcoding approaches for marine benthic biodiversity monitoring. The studies in this thesis are grouped according to two purposes; the first purpose was to compare metabarcoding methods to classic morphological methods. In the methodological studies (Chapters 2 and 3), the Wadden Sea intertidal ecosystem is used as a model system. The second purpose was to apply the metabarcoding methods in ecological field studies towards anthropogenic impacts on marine benthic biodiversity in different areas: the intertidal Wadden Sea (Chapter 4), The North Sea (Chapter 5), and the deep- sea (Chapter 6).
Chapter 2 compares the effectiveness of different workflows used in the metabarcoding approach. Three DNA extraction methods were compared to the classic morphological approach and a comparison was made between two taxonomic assignment methods. Whereas chapter 2 is focused on qualitative data, the application of metabarcoding approaches to assess marine benthic biodiversity in a quantitative manner is assessed in chapter 3. This chapter aims to examine to which extent abundance estimates based on metabarcoding results, either via a frequency of occurrence approach or via a relative read abundance approach, are comparable to abundance estimates based on the morphological approach.
The results derived from chapter 2 and 3 were used as the base for applying metabarcoding methods to infer the benthic community in different ecosystems, possibly subjected to anthropogenic influences, in three case studies. The three marine benthic ecosystems were all soft-sediment ecosystems but highly varied in complexity. Chapter 4 describes the influence of bottom dredging for the lugworm Arenicola spp. in the intertidal Dutch Wadden sea on the benthic biodiversity there. In this ecosystem, species diversity is relatively low and is well studied with the traditional morphological approach. Therefore, conclusions derived from the metabarcoding approach could be compared to those from the morphological approach. Chapter 5 examines the impact of a gas platform on the benthic community located on the continental shelf in the Southern North Sea. This ecosystem is species rich and moderately studied. Spatial variations within the benthic community, possibly induced by impacts of the gas platform, were examined both with a metabarcoding and morphological approach. In the last case study, chapter 6, the metabarcoding approach was used to explore soft-sediment deep-sea communities in the background of the Rainbow hydrothermal vent field at the Mid-Atlantic Ridge. These so called
‘background communities’ are relatively unknown but could potentially suffer loss of biodiversity due to deep sea mining. This study examined patterns in the benthic communities and tries to correlate these to influences of hydrothermal plumes. Also, an attempt was made to detect pelagic larvae of benthic species at different locations around the hydrothermal vent field with the metabarcoding approach.
